KBAI Lesson 4 -Frames


Frames are powerfull knowledge representation technique in AI field.

###Functions of Frames Functions of frames were explained using the intrepretation of sentences. Based on a verb in the sentence a representation for the sentence is built based on which the sentence is intrepreted. Frames are framed based on verbs or nouns. Linked Frames are created to link between nouns in the sentences.

###Properties of Frames

  • They are mostly stereotypes - i.e generic which depends on persons or culture specific
  • It provides default values-eg frog eat its not alive. Similar to constructor in OOPS
  • Exhibit inheritance- similar to OOPS concepts

###Relationship ####Frames and Semantic networks Frames and semantic networks are closely related. In order to solve Raven’s problem we will be building frames for objects in the figure. All the relationships in semantic networks are captured in frames using variable names

####Frames and Production Systems In production Systems we have semantic, procedural and episodic . For example in sentence intrepretation when parsed to production systems the template of frames are pulled from semantic portion based on the verb in the working memory. Once the frames details are pulled

###Frames for advanced Sense Making For complex sentences which had dual meanings like the earthquakes example provided in the lecture frames may not work properly. Since the stereotypes it may not work exactly as required for the given problem. This will be addressed by common sense problems.

###Cogntive Connection

  • Frames are structured knowledge representation. Production systems are atoms whereas frames are like molecules. It can capture lot of information in an organized way.
  • It helps in top down approach. It allows in retrieval of information from memory.
  • It helps in stereotypes identifications. Which is cognitively efficient.